Challenges with WPP

In this chapter some of the challenges and disadvantages of wind power production are de- scribed. Especially the handling of the production- and consumer balance is discussed. Passive producers, such as run-of-river hydro and wind power have a disadvantage in the two-price system. As these are intermittent resources the production is only controlled by the wind speed and inflow of water in the hour of operation. Uncertain forecast gives significant imbalances. The deadline for changing the schedule is 45 minutes before the hour of operation. For the run-of-river hydro this may be acceptable, as the inflow does not change significantly over 45 minutes. Furthermore the prognosis for inflow is well developed as it makes a great di↵erence in optimization of traditional hydro power plants. Therefore the production of run-of-river hydro is more predictable than the production of wind power plants. The run-of-river power plants normally has a high output when there is a lower demand, typically in the spring and autumn. In this period the prices are normally low

and the price of balancing correlates with this. The wind speed, however, may change significantly in only five minutes. In a larger area the short term variations will be smoothed out. In Denmark the maximum hourly production change is approximately 15-20% [1]. This represents a risk as an imbalance of 20% at an unfavorable time can represent a great challenge when handling the balances.

Furthermore, there are speculations that a major investment in intermittent renewable en- ergy sources will lead to higher fluctuations in both the spot and balancing prices. Increased volatility in the power markets increases the risk for economic losses in the production im- balance. For this reason it may be unfavorable for a producer to have large penetration of wind in the production portfolio, compared to flexible hydropower.

Today the production schedule for wind power is established on the basis of future projections of the wind speed, provided by the weather services Storm.no and the Norwegian Metrological Institute. These have proven to di↵er from actual wind speed. As mentioned above, the production schedule can be changed 45 minutes before the hour of operation. An alternative method for the wind forecasting has been developed and proven to be better than the current projections. In this alternative method the production of the next hour is scheduled as the measured production this hour. The production scheme is changed and sent to the TSO 45 minutes before the hour of operation. This method is hereafter referred to as the re-bidding procedure. An example of the two procedures is shown in Fig. 2.14.

Figure 2.14: Actual production, projection and the re-bidding procedure. Bessaker wind farm - July 8th, 2012 [29].

The red line represents the production reported to Nord Pool Spot according to weather forecasts and the blue line is the actual production. The green line is the production reported one hour ahead and set equal to the production in the current hour. We see that the

green line is more accurate then the red line. The purple line represents the change of schedule in each hour. We see that the modified method matches the production profile better than the prognosis. The total change of schedule in this example is 73.65 [MWh] [29]. The forecast error over 24 hours is hardly conclusive. This example is included to get a better understanding of how TrønderEnergi handle the imbalances caused by Bessaker wind farm.

For a wind power producer it is desirable to have the imbalance priced after the one-price model. In this model the producer has the possibility to gain a profit on their uncontrollable imbalance. Therefore it is beneficial to have most of the imbalance in the consumer balance shown in Eq. 4.15. In this equation the term trade prior to the operation hour is the trade done in the Elspot and Elbas market witch Nord Pool Spot operates. In the spot market the trade is done 12:00 CET the day before. This scheme is set up according to weather forecasts provided by Storm.no and the Norwegian Metrological Institute. The term trade before operating hour can be changed intraday by using Elbas, but this is not considered in this section. By changing the planned production for the next hour, the production imbalance is hopefully reduced, as the modifications done normally are more accurate than the weather forecasts.

When changing the production schedule 45 minutes before the hour of operation the produc- ers change the planned production for the forthcoming hour, reducing the di↵erence between actual and planned production. This will reduce the imbalance in the production balance in equation 4.14. The original schedule is delivered for each hour before 12:00 CET the day ahead, and is according to di↵erent weather forecasts. Furthermore, the di↵erence between planned production and trade prior to the operation hour will increase when changing the planned production, so the consumer imbalance in Eq. 4.15 is increased. In this manner some imbalance is moved from the production balance to the consumer balance. This is as mentioned desirable as the consumer balance is priced after the one-price model, while the production imbalance is priced after the two-price model. An example of how the bal- ance sheet would be changed using the re-bidding procedure is found in section 2.2.3, and illustrated in Table 2.4 and Table 2.5.

Arne Kjell Nystad [24] in Statnett points out that increased integration of wind power will a↵ect the time horizon in the balancing market. Nowadays it is the consumer gradient, not the production, that varies during the hour of operation. This will probably change with increased integration of wind power. With a more unstable production it will be harder to manage the frequency control, the balancing of the production and consumption. This will increase the need for regulations with a short time horizon, which will give unfavorable results for producers. In the active balancing market, the producers assume that if they are put into operation the start and stop costs will be covered when they get their demanded price in the hour of balancing. The TSO, however, have the right to activate the bids at any time. With a more uncertain production portfolio in the system, some producers may be activated for only 10-15 minutes. In such situations they will see a loss. For the next bid of balancing power, a reasonable producer will increase their bid. This, combined with

increasing demand for balancing power, may contribute to a increasing di↵erence between the spot price and balancing price.

Chapter 3

Analysis of the Planning Situation

In this section di↵erent variables in handling the uncontrollable wind power production are analyzed. In Section 3.1 a case study of the current handling of Bessaker wind farm is done and in Section 3.2 the future development of the balancing market is described. The opportunities for a integrated balancing market using cross-border transmission capacities is evaluated in Section 3.2.1. In Section 3.3, the development of the intraday market in Norway is discussed, followed by Section 3.4 which describes the benefits and drawbacks of using internal hydro power to compensate the balance position. There is not much literature on the two latter subjects, so conversations with actors in the market have been a good source of information. Finally, in Section 3.5, the development of improved weather prediction using numerical weather predictions are elaborated.